Importance Weighting in Subjective Well-Being Measures: Using Marginal Utilities as Weights for Domain Satisfaction

Whether perceived importance can be used as a weighting factor for the aggregation of domain satisfaction is an important issue in the literature on subjective well-being (SWB) measures with multidimensional domains. This paper extends the weighting approach proposed by Benjamin et al. (Am Econ Rev 104:2698–2735, 2014) and uses it to weight domain satisfaction. The weights estimated by this approach can be interpreted as marginal utility, and the weighted average of the domain satisfaction of each individual can be interpreted as the individual’s utility. Using a data set collected from a 2015 survey that we conducted with residents of all Japanese prefectures, we show that this weighted average improves the goodness of fit for overall SWB measures compared with the unweighted averages of domain satisfaction. This finding supports incorporating perceived importance measures into SWB measures. Moreover, we show that, although there are differences in perceived importance among some domains that are less relevant to specific sub-populations, such as work for people above 65 years or relations with family for those who live alone, the relative importance of most domains is similar across the different categories of gender, household income, age, and household composition. Allowing for heterogeneity in the importance weights across these demographic groups provides only negligible improvements to the validity of the domain importance. The approach suggested in this paper can be considered a resident-oriented approach in the sense that the weights are derived from the perceived importance of residents themselves and are thus expected to contribute to resident-oriented policymaking.